by way of the highly selective tri-nbutyl thiophosphate extraction. [Of 35 diverse elements studied, only silver and mercury are extracted efficiently ( 6 ) . ] The behavior of palladium in the analytical scheme was not examined; however, since this element has never been detected in searvater, interference frolrl tllis appears unlikely. rl1 the colorinletric method using p-dilllethylaniinobenzalrhodanine, interferences are renioved by coprecipitation of silver with tclluriuni (8). hforeover, the consistency of the analytical determinations for different aliquot volumes as well as for the two separate quant’itative procedures points to the reliability of the method. The separation of silver from seawater by cocrystallizntion &h tllionalid offers advantages Over ‘lore tionnl methods. The crystallization is
perfornled simply, and the crystals are of such nature as to permit rapid filtration. In the process, the major fraction of silver (>go%) is separated from the large bulk of inorganic constituents of seawater with the use of a moderate quantity of carrier. The ultimate iS0lation is SinlPlified by the ability to remove the organic carrier through acid Oxidation. These factors contributed to ease and reliability in the determination. LITERATURE CITED
(1) Black, W. A. P., Mitchell, It. L., J. h f a r k e Bid. ASSOC. U . K . 30, 575
(1952).
(2) Cave, G. C. B., Hume, D.
s.3 A N A L .
CHEM.24, 1503 (1952). (3) Fajans, K., Berr, p., Be?. deut. them Ges. 46, 3486 (1913). (4)Haber, F., 2. angew. Chenz. 40, 303 (1927).
(5) Hahn, O., ivaturwissenschaften 14, (6)‘lg6 Handle?;, (19261.T. H., Dean, J. A., ANAL. cHEM, 32, 1878 (1960). (7) Noddack, I., Noddack, W.,Arkiv Z O O l . 32A1 1 (1939). (8)ANAL. Sandell, E’23, 1863 KeumayerJ CHEM. (1951). J ’ J.p (9) Vnderwood, 8. L., Burrill, A. XI., Rogers, L. B., Ibid., 24, 1597 (1952). (10) Jvahl, A. c.,Banner, x. A , , “ h d i o activity Applied to Chemistry,” Wile?;, New York, 1951. (11) ‘veiss, H. \T,, ~ ~ M,i G, , , L4raL, CHEM.32, 475 (1960). (12) Weiss, H. Y.) Lai, M. G.: J . Inorg. h’?tcl. Che7n. 17, 366 (1961). (13) Weiss, H. V., Lai, M. G., Talanta 8, 72 (1961). (14) Veies, H. V., Lai, M. G., Gillespie, A., Anal. Chim. Acta. 25, 550 (1961i . (15) ~veiss, He V., Shipman, H., - 4 S A L . CHEM. 34, 1010 (1962). (16) \\r&her, F. J., “Organic Analytical Reagents,” Vol. IV, p. 165, D. Van Yostrand Co., Inc., Sew York, 1948. B‘j
RECEIVEDfor review March 15, 1962. dccepted hIay 11, 1962.
Quantitative Determination of Ferrous Oxide in the Presence of MetaIIic Iron and Ferric Oxide in Difficultly-Soluble Minerals, Treated Ores, and Slags M. G. HABASHY Special Research laboratory, 25 Sultan Hussein St., Alexandria, Egypt, U.A.R.
b Ferrous oxide is determined in the presence of metallic iron and ferric oxide in difficultly-soluble samples such as ilminite, siliceous magnetite ores, treated haematite, slags, and minerals of pyritic origin. A specially designed apparatus is described in which samples are ignited in a limited measurable volume of oxygen gas or purified air. The volume of oxygen gas consumed in converting ferrous oxide to ferric oxide i s measured, modulated to STP, and converted to ferrous oxide content. The interference from many elements or compounds which absorb oxygen can b e eliminated. Results show a mean tol0.1 %. erance of
*
for the determination of FeO content in difficulty-soluble samples which contain sulfur, e.g., pyritic sulfur, are not adequate. illso the alkali carbonate fusion decomposition of n n acid-insoluble sample does not give a true picture of the initial FeO content. even when COS is used to exclude air continuously throughout the process. I t was found t h a t results differed among samples by 15 to 20% when the samples were decomposed by the potassium sodium carbonate fusion ETHODS
and by the pressurized HCl method described by Pina (4). These differences might have been due to the presence of metallic iron (the sample was refining iron scales from the Off0 furnace) in both reactions, or might have been due to the presence of acid salts in the sample which affected the alkali fusion. However, not all acid-insoluble samples, e.g., ilmenite and copper slag samples, can be dissolved by the pressurized HC1 solution method (6). Dittrich and Leonhard ( 2 ) attempted to dissolve samples in H2S04and HF, and Shein (7) used a mixture of H3P04, and H ~ S O I ;and V20sas a solvent with a stream of C 0 2 over a water bath, but neither method completely decomposed the above mentioned samples. Shein ( 6 ) ignited the sample under a current of 0 2 a t 1000” C. for 30 minutes; COZ and the water of constitution were absorbed b y Ascarite and weighed. The ignited sample was weighed and the per cent FeO xvas obtained by calculation. Results were accurate to 0.1%. Later Shein ( 7 ) adapted this method t o include magnctite and other difficultly-soluble samples by igniting the sample in nitrogen to expel volatile matter before ignition n ith 02. Shein assumed that carbon was present only as
a i d he neglected the possible effect of Y or other elements, e.g., metallic Fe which could be present in slag samples and treated ores. Also he neglected the FeO present as 2 Fe0.SiO2. This compound cannot react with oxygen a t 1100” C. as shown b y Plotkin (6). Shein’s method was adapted on a semimicro scale b y de U7et (8)who determined FeO in chromite with an accuracy which was within 0.05 to 0.1% of the total analysis. This paper describes the determination of per cent FeO in samples which might contain S, C, metallic impurities, moisture, water of constitution, or a ferrous /silica complex. EXPERIMENTAL
The technique followed by the author necessitated the construction of a special apparatus, hereafter referred to as the ferrometer, which is shown in Figure 1. Samples. T h e following samplcs were used: Ilmenite sample, Srttional liesearch Center, E l Dokky, Cairo Mazut-reduction spongy iron sample, prepared by the author from Aswan haematite without using coke (a special method). VOL. 34, NO. 8, JULY 1962
1015
Raw dolomite rock sample, Egyptian Dolomite Mines, Cairo. Copper slag sample, Egyptian Copper Works, Alexandria. Open hearth slag sample, Egyptian Copper Works, Alexandria. Calcium carbonate sample (AnalaR), German type (Merck) Ferrosoferric oxide sample, prepared b y electrolyzing redistilled water be-
tween pure iron electrodes for 20 days and dried a t 150" C. Calibration of Apparatus : Electrolytically prepared ferrosoferric oxide is used. It is easily soluble in HCl a n d its FeO content can be checked using t h e potassium dichromate method. A blank test r u n on the apparatus should give zero results. Procedure. Drive the solution in
d
m 5
!r 13
Figure 1. 1.
i i
"B u 18
$7'
I: u 15
Diagram of apparatus
Series o f gas-washing bottles containing Animal charcoal and glass wool b. Saturated solution of potassium permanganate c. 30% potassium hydroxide solution d. 10% sulfuric acid solution e. Distilled water f. Empty flask g. Silica gel (indicator type) h. Gloss wool 2. Three-way glass tap 3. 300-ml. (internal capacity) condenser; outlet of jacket i s fitted with 0-45' C.thermometer 4. 50-ml. buret 5. Reservoir (used as atmospheric pressure equalizing vessel and contains acidulated water made of 30% sodium chloride solution, 10% sulfuric acid, and methyl red indicator) 6. Pyrex combustion tube (90cm.; porcelain tube gave same results) 7. Electric heater automatically controlled by an electronic thermoregulator using a sensitive thermo. couple, e.m.f. of which is amplified by an electronic amplifier 8. Electromagnetic ring surrounding combustion tube to permit movement of combustion boat 9 with the aid of internal iron core 20 and platinum wire P inside the tube without opening the latter during the determination. Magnet i s composed of 400,000 turns of enameled copper wire of 0.1-mm. diameter which ius' slides over the combustion tube. Supplied with a direct current voltage of 90 volts and condensed on a high voltage condenser of 1 00-Ff. capacity 9. Porcelain combustion boat 10. Thermocouple, platinum-platinum alloy 1 1. Two-way glass tap leading to U-tubes, 13, 15, and 18 12. Two-way tap leading to bubbler o f acidulated water U-tube 13 13. Pyrex U-tube containing acidulated water made as described 14. Two-way glass tap leading to alkali U-tube 15 15. U-tube containing 30% potassium hydroxide solution 16,16', and 16". Trap made of funnel filled with soda lime and glass wool. 17. Aluminum jacket surrounding cold part of combustion tube. When compressed conditioned air is passed through this jacket, temperature of cold region of combustion tube can be controlled to reach Upper (horizontal) half, 178 room temperature. A 0-45' thermometer is fitted through the jacket. of jacket is movable to permit working o f electromagnet 8 b y sliding it aver combustion tube 18. Pyrex U-tube containing paraffin oil 19. Two-way glass tap leading to paraffin tube 18 20. Internal iron core of electromagnet 8 a.
1016
*
ANALYTICAL CHEMISTRY
U-tubes 13, 15, and 18 (Figure 1) into the left-hand side limbs by lowering reservoir 5 until t h e solutions in the U-tubes reach marks m2,ma, and m4, respectively, while taps 2 and 11 are i n a position connecting condenser 3 with tube 6 and the U-tubes (each in turn). Close tap 11. Connect taps 12, 14, and 19 so as to permit the entrance of the gas to the U-tubes through the long tube L . Return of the gas should be via the short tubes 1. (Automatically adjusting floating glass valve systems can be used instead of the taps.) Accurately weigh about 2 grams of sample (according to the FeO content) into the combustion boat and place in the right-hand side of the combustion tube after connecting the side ring of the boat with the platinum wire P which is sealed from the other end to an iron core that functions as a driver from the effect of the outer electromagnet 8. Close tube 6 (the boat is still in the cold region of the tube) and apply heat until the temperature is stabilized a t 900" C. Two large burners give satisfactory results if there are no air currents in the laboratory. Connect condenser 3 to the atmosphere through the opening 0. Raise reservoir 5 until the acidulated water fills condenser 3 to mark ml. Operate tap 2 to connect washing bottles 1 with condenser 3, then lower reservoir 5 to fill condenser 3 with clean air until the level of the acidulated water reaches a certain reading on the lower end of buret 4. The abore three steps mag be repeated. Turn tap 2 to connect condenser 3 with tube 6 and record the buret reading. Move electromagnetic ring 8 over the tube to drive the boat to the heated region, cut off the current, and slide the ring back to the right end of the combustion tube. (The iron core of the magnet which is inside the tube is far away from the heated region but as an additional precaution it may be gold plated.) Open taps 11 and 12 and raise reservoir 5 until the acidulated mater reaches mark ml and then lower the reservoir back t o its original position. On lowering, the acidulated water of C-tube 13 must exactly reset to mark m2. Close taps 11 and 12. Read the buret while equalizing the inside pressure with that of the atmosphere by raising reservoir 5 until the two acidulated water levels inside buret 4 and reservoir are a t the same level. Purify the gas by passing i t through the K O H U-tube 15 by repeating the above steps but using taps 11 and 14 instead of 11 and 12 to avoid the interference of carbonate COz from the sample. Pass the igniting gas (air) over the sample until two readings of the buret are similar (usually four times are adequate). Take the reading after leaving the gas inside the condenser for 3 to 5 minutes to reach the starting thermal equilibrium throughout the different parts of the apparatus. De-
duct the portion of oxygen consumed to ignite the sample from the difference between the initial buret reading and the final reading. If free sulfur is suspected as in pyritic or other sulfide ores, carry out the determination using paraffin oil instead of acidulated water inside reservoir 5 and use the third U-tube 18 filled with paraffin oil. Allow sufficient draining time before taking the reading when using paraffin oil because of its high viscosity (or mix a suitable nonvolatile thinner with the paraffin). The difference between the volume of air used for ignition when the paraffin tube is used and t h a t after this volume is passed through the acidulated water U-tube is the volume of SOz from which the S content of the sample can be calculated. If the presence of free carbon is suspected in the sample, e.g., from sintering processes of iron with coke or mazutreduction preparation of spongy iron (by the author), determine the carbonate carbon first using air-free nitrogen instead of air through the abore .teps. I n other samples, determine the total carbon by repeating the determination with air as in the above steps. Hy calculation the difference will indicate the free carbon content from which the equivalent volume of oxygen consumed can be calculated. Convert the volume of oxygen absorbed by the FeO sample to STP, then multidv b "v 1.283 if the taken weight of the sample is 1 gram to give the i e r (cent FeO. I
"
TREATMENT OF INTERFERENCES
When sulfur is ignited in oxygen, no change in volume occurs. T h a t is, when the sample, containing sulfide sulfur, is ignited in the ferrometer, no change in volume is expected from the ignition of the sulfide radical. This fact is restricted to those samples which contain no basic oxides such as CaO, MgO, and MnO. The presence of such
Table
II.
oxides will, more or less, fix the sulfide sulfur u p t o 900" C. as sulfates, during ignition in oxygen (or air). Although FeO is also a basic oxide, i t does not
Table 1.
Quantitative Analysis of Selected Testing Samples"
Sample Numbers 4 5 OpenReduced Hearth HaemaRaw Copper Iron Percentage Ilmenite tite Dolomite Slags Slags hioisture 0.40 < O . 05 0.39 0.58 0.33 Si02 4.04 2.30 1.40 12 90 10.30 ZnO Nil Xi1 Xi1 3 15 0.65 A1203 1.70 10.60 0.57 5.84 2.30 22.75 19.95 Fen03 0.18 21 35 12,54 Fe (metallic) Nil Nil 41.30 3.90 5.80 - FeO detailed in followi ng Tables I1 and I11 Ti02 41.15 Nil Nil Si1 Traces PbOz Nil Nil Nil 6 91 0.25 SnOz Nil Xi1 Nil Nil 2.35 NiO Nil Nil Si1 2.40 4 78 MnO? 0.25 0.40
